The changing face of the Mediterranean – Land cover,demography and environmental change: Introduction and overview

Pollen archives

Fossil pollen archives from sediment cores have long been a key resource for localised landscape and land cover reconstruction, although broad-scale studies have often been limited by dating imprecision and data coverage. Over the past two decades, methods have become increasingly sophisticated and greater efforts have concentrated on improving the quality of large databases and their chronological precision (e.g. Giesecke et al., 2014). Consequently, the role of pollen archives has expanded, as various efforts have concentrated on providing multi-archive, increasingly large-scale land cover narratives via a variety of statistical methods. In Europe, the development of a shared fossil pollen database (European Pollen Database, Fyfe et al., 2009; Leydet, 2007–2013) and modern pollen database (Davis et al., 2013), and pan-regionally relevant pollen indices constitute important advances (Fyfe et al., 2018; Woodbridge et al., 2018). Land cover reconstruction methods such as pseudo-biomisation (Fyfe et al., 2010), REVEALS (Sugita, 2007; Trondman et al., 2015) or plant functional types (Davis et al., 2015) have also provided important new approaches, with good overall agreement for temperate and boreal Europe in terms of when and where forest cover was cleared to make way for fields and pastures (Roberts et al., 2018). While pollen analysis is less easily applied in some very dry areas of the Mediterranean (e.g. the Libyan Desert), it remains the most widely available means of reconstructing past Mediterranean land cover, and the promising start at integrated land cover assessments in more northerly regions offers the possibility that these or similar approaches might be used for reconstructing the history of Mediterranean landscapes as well.

That said, the pollen taxa that provide clear indicators of anthropic activity in temperate Europe may not carry the same meaning for Mediterranean landscapes, so there has been a pressing need to establish a new, different set of land cover classes that is more relevant to the latter region’s vegetation communities (such as maquis-garrigue or olive groves: Fyfe et al. 2015). For this and in the regional papers that follow, we have used cluster analysis of all modern and fossil Mediterranean pollen samples to generate 16 coherent groupings of plant taxa (Fyfe et al., 2018; Woodbridge et al., 2018), while also exploring the insights provided by a number of pollen indices that group together related taxa, such as tree crops (Olea, Juglans, Castanea or OJC, sometimes also joined by Vitis, OJCV). Most of the pollen data used by subsequent papers in this special issue come from 177 sequences archived in the European Pollen Database (EPD; Leydet, 2007–2013), for which good age-depth chronologies have already been established (Giesecke et al., 2014). For specific case study regions, these have been supplemented by 76 additional records (listed individually for each case study region paper) that have both been harmonised taxonomically and, where necessary, given new or reassessed age-depth models. This resulted in a total of 253 fossil pollen records summed into 200-year time windows for the duration of the Holocene and 1798 modern pollen sites.

Radiocarbon dates and settlement surveys

One major academic shift in archaeology over the past decade has been an increased willingness to address the crucial, but enduringly difficult, challenge of reconstructing pre- and proto-historic population trends in a statistically useful way. Two methods that have been increasingly prioritised involve collection and analysis of cross-regional aggregations of (a) radiocarbon dates and (b) settlement surveys. The first of these approaches usually involves the summing of large groups of anthropogenic radiocarbon dates, under the assumption that a sufficiently big sample will exhibit highs and lows in date frequency in step with varying intensity of human activity through time, and hence can be thought of as a high chronological resolution proxy for long-term population dynamics. While the basic technique has a deeper pedigree going back three decades (Rick, 1987), it has seen far greater emphasis over the past 5–10 years, especially with regard to more refined statistical modelling and testing. For example, the approach has been successfully applied to later prehistoric Europe north of the Alps (Shennan et al., 2013), revealing striking patterns of demographic boom associated with the first Neolithic farming communities, sometimes followed by subsequent population collapse. Further comparison of these summed radiocarbon results with aggregated pollen records has revealed what, on the whole, are excellent matches between the changes exhibited by these two kinds of archive through time (British Isles: Woodbridge et al., 2014, Southern Germany: Lechterbeck et al., 2014). This kind of cross-disciplinary synthesis offers huge potential, we would argue, for understanding whole socio-ecological trajectories of Holocene population and land cover change.

A second popular demographic proxy involves systematic synthesis of what we know about past settlement from archaeological surface survey projects and excavations. Archaeological sites have been mapped, dated and counted for many years, but what has changed more recently has been the availability of larger datasets deriving from multiple sources (both intensive systematic surveys and more extensive reconnaissance or site gazetteers), better statistical handling of geographic biases and greater attention to the modelling of chronological uncertainty (e.g. Palmisano et al., 2017). Wherever available, the basic information collected in this manner comprises settlement numbers, locations and sizes, with each site’s traditional attribution to a particular cultural period (of varying length) converted to approximate absolute chronology so that the results are most closely comparable with the summed radiocarbon probability distributions.

As with pollen samples, numbers are similarly large for archaeological data, with > 20,000 radiocarbon dates from excavated Mediterranean archaeological sites, and tens of thousands of settlements recorded by systematic regional survey datasets (~30,000). The analysis of these large-scale spatial and temporal datasets has necessarily also involved a combination of geographic information systems (GIS), multivariate generalisation, Monte Carlo simulation and probability-based statistical techniques, as well as considerable effort at data harmonisation.

Proxies for climate, fire and erosion

Beyond the core evidence contributed by palynology and archaeology (discussed above), this special issue has also provided an opportunity to elicit up-to-date reconstructions of Mediterranean hydro-climate, fire and erosion regimes. While pollen data have been used by some authors to reconstruct Holocene climate changes in the Mediterranean (e.g. Guiot and Kaniewski, 2015), there is a risk of circular reasoning in using the same source of information to establish both potential cause (climate) and consequence (land cover). Fortunately, there exist proxy records of Mediterranean climatic change from data sources other than pollen, such as stalagmite and lake sediment stable isotopes, which are amenable to synthesis and statistical transformation (Finné and et al., 2011; Labuhn et al., 2018; Luterbacher et al., 2012). This climatic evidence needs to be critically evaluated alongside human agency in the interpretation of results concerning land cover change. Another significant agency in the summer-dry Mediterranean is wildfire, which can be influenced by both climate and human activities. Long-term trends in fire frequency and intensity can be revealed through charcoal records, especially from micro-charcoal particles that are often identified and counted on pollen slides (Vannière et al., 2011). One of the main consequences of land cover change has been altered erosion and hydrological regimes, which can be reconstructed from geomorphological records of sediment flux and river flooding/sedimentation. There are good proxy data for these from Mediterranean valleys and lake basins (e.g. Benito et al., 2015; Dusar and et al., 2011; Macklin and Woodward, 2009), some of which have been subject to statistical methods similar to those used in archaeology (e.g. on aggregations of radiocarbon dates).

Regional case studies

Palmisano et al. (this issue) consider the south-central Levantine region, in many ways the precocious child among these regional cases studies with the impact of such things as equids, sailing ships, olive and vine cultivation, urbanism all occurring earlier than in other study regions, as well as a particularly dense settlement record to make use of. Woodbridge et al. (this issue) address a portion of southern Anatolia, which brings together a strong set of pollen archives, considerable wider environmental work and several archaeological surveys, in an area that is particularly relevant for understanding how farming moved into Europe. Weiberg et al. (this issue) consider Greece over a period spanning the earliest European Neolithic colonisations, the appearance of Europe’s first palatial, Bronze Age societies and the emergence of Classical Greece, with opportunities for comparing northern and south Greek trajectories across a balanced set of settlement, radiocarbon and pollen evidence. Stoddart et al. (this issue) consider the crucible of cultural change in central western Italy, with a scope that stretches back to the later Mesolithic as in other papers, but with a forward eye on dynamics that would eventually produce Rome, a pre-modern city of perhaps a million inhabitants and the heart of an empire whose environmental footprint was enormous. Berger et al. (this issue) analyse an unusually data-rich case study region in southern France with its large sets of radiocarbon dates and many pollen archives, as well as a mix of settlement data from developer-led rescue archaeology, research-led intensive field surveys and sub-surface information from trainline transects. Fyfe et al.’s study (this issue) of eastern Spanish patterns is blessed by a good set of pollen archives, and an opportunity to compare mainland coastal environment trajectories with those pertaining in the more belatedly colonised habitats of the Balearic Islands.

In contrast to these six case studies from the northern coasts of the Mediterranean, one of the most widely agreed deficiencies in our current understanding of Mediterranean cultural and environmental dynamics relates to the far more limited evidence currently available for the southern side of the sea (e.g. Broodbank, 2013: 36–39), with a dearth of pollen studies from most of North Africa, very little radiocarbon sampling from Holocene archaeological sites and patchier archaeological investigation outside Egypt. Such problems can only truly be solved by intensified attention and better-funded work in these southern regions, but it is nevertheless very welcome to be able to add a seventh case study from Morocco that offers a more tentative but important first impression (Cheddadi et al. this issue).

Pan-Mediterranean challenges

Beyond these regional case studies, other Issue contributors take the chance to address some key pan-Mediterranean challenges whose results intersect in important ways with the demographic and land cover evidence. Mercuri et al. (this issue) grapple with the methodological challenges of synthesising on-site palynological and anthracological data and thereby provide a much-needed counterpoint to the evidence from off-site Mediterranean archives. Finné et al. (this issue) likewise contribute a significant methodological scaffold using standardised scores to combine Mediterranean palaeoclimatological records, with results that are not only interesting in their own right but also better support ongoing comparative work and which reveal significant regional differences in climate histories. Walsh et al. (this issue) summarise and interrogate trends in erosion as preserved in Mediterranean sedimentary archives, thereby providing an important and complementary view of landscape change to land cover focused perspectives of palynology. Connor et al. (this issue) explore fire regimes as manifest in the natural and anthropogenic charcoal records of the Mediterranean, specifically in Iberia, with obvious wider relevance in terms of how changing fire intensities correlate with climate change and/or human action. Langgut et al. (this issue) bring together what we currently know about the chronology and geographical patterning of olive domestication, a key economically important tree crop, and ensuing cultivation across the entire basin, with fresh focus on palynological evidence, but also due attention to how this compares with genetic studies and archaeobotanical remains.

In a final paper, Roberts et al. (this issue) step back and offer a first reflection on what kind of joined-up story all these contributions might offer, with a view also to future research directions. Without anticipating the results of the individual studies that follow, one cross-cutting feature of these studies is the way they provide complementary perspectives on Mediterranean landscapes being shaped by increasingly coupled, natural and anthropogenic regimes as the Holocene progressed. More generally, the fact that longitudinal results from Mediterranean environmental science, archaeology and landscape history can be juxtaposed as they are in this issue, in both consensual and adversarial ways, bodes extremely well for future efforts to describe the changing face of this ecologically striking and culturally consequential part of the world.